Matching Items (58)
Filtering by
- All Subjects: Electrical Engineering
- Creators: Ogras, Umit Y.
Description
The drive towards device scaling and large output power in millimeter and sub-millimeter wave power amplifiers results in a highly non-linear, out-of-equilibrium charge transport regime. Particle-based Full Band Monte Carlo device simulators allow an accurate description of this carrier dynamics at the nanoscale. This work initially compares GaN high electron mobility transistors (HEMTs) based on the established Ga-face technology and the emerging N-face technology, through a modeling approach that allows a fair comparison, indicating that the N-face devices exhibit improved performance with respect to Ga-face ones due to the natural back-barrier confinement that mitigates short-channel-effects. An investigation is then carried out on the minimum aspect ratio (i.e. gate length to gate-to-channel-distance ratio) that limits short channel effects in ultra-scaled GaN and InP HEMTs, indicating that this value in GaN devices is 15 while in InP devices is 7.5. This difference is believed to be related to the different dielectric properties of the two materials, and the corresponding different electric field distributions. The dielectric effects of the passivation layer in millimeter-wave, high-power GaN HEMTs are also investigated, finding that the effective gate length is increased by fringing capacitances, enhanced by the dielectrics in regions adjacent to the gate for layers thicker than 5 nm, strongly affecting the frequency performance of deep sub-micron devices. Lastly, efficient Full Band Monte Carlo particle-based device simulations of the large-signal performance of mm-wave transistor power amplifiers with high-Q matching networks are reported for the first time. In particular, a CellularMonte Carlo (CMC) code is self-consistently coupled with a Harmonic Balance (HB) frequency domain circuit solver. Due to the iterative nature of the HB algorithm, this simulation approach is possible only due to the computational efficiency of the CMC, which uses pre-computed scattering tables. On the other hand, HB allows the direct simulation of the steady-state behavior of circuits with long transient time. This work provides an accurate and efficient tool for the device early-stage design, which allows a computerbased performance evaluation in lieu of the extremely time-consuming and expensive iterations of prototyping and experimental large-signal characterization.
ContributorsGuerra, Diego (Author) / Saraniti, Marco (Thesis advisor) / Ferry, David K. (Committee member) / Goodnick, Stephen M (Committee member) / Ozev, Sule (Committee member) / Arizona State University (Publisher)
Created2011
Description
In semiconductor physics, many properties or phenomena of materials can be brought to light through certain changes in the materials. Having a tool to define new material properties so as to highlight certain phenomena greatly increases the ability to understand that phenomena. The generalized Monte Carlo tool allows the user to do that by keeping every parameter used to define a material, within the non-parabolic band approximation, a variable in the control of the user. A material is defined by defining its valleys, energies, valley effective masses and their directions. The types of scattering to be included can also be chosen. The non-parabolic band structure model is used. With the deployment of the generalized Monte Carlo tool onto www.nanoHUB.org the tool will be available to users around the world. This makes it a very useful educational tool that can be incorporated into curriculums. The tool is integrated with Rappture, to allow user-friendly access of the tool. The user can freely define a material in an easy systematic way without having to worry about the coding involved. The output results are automatically graphed and since the code incorporates an analytic band structure model, it is relatively fast. The versatility of the tool has been investigated and has produced results closely matching the experimental values for some common materials. The tool has been uploaded onto www.nanoHUB.org by integrating it with the Rappture interface. By using Rappture as the user interface, one can easily make changes to the current parameter sets to obtain even more accurate results.
ContributorsHathwar, Raghuraj (Author) / Vasileska, Dragica (Thesis advisor) / Goodnick, Stephen M (Committee member) / Saraniti, Marco (Committee member) / Arizona State University (Publisher)
Created2011
Description
This work is focused on modeling the reliability concerns in GaN HEMT technology. The two main reliability concerns in GaN HEMTs are electromechanical coupling and current collapse. A theoretical model was developed to model the piezoelectric polarization charge dependence on the applied gate voltage. As the sheet electron density in the channel increases, the influence of electromechanical coupling reduces as the electric field in the comprising layers reduces. A Monte Carlo device simulator that implements the theoretical model was developed to model the transport in GaN HEMTs. It is observed that with the coupled formulation, the drain current degradation in the device varies from 2%-18% depending on the gate voltage. Degradation reduces with the increase in the gate voltage due to the increase in the electron gas density in the channel. The output and transfer characteristics match very well with the experimental data. An electro-thermal device simulator was developed coupling the Monte Caro-Poisson solver with the energy balance solver for acoustic and optical phonons. An output current degradation of around 2-3 % at a drain voltage of 5V due to self-heating was observed. It was also observed that the electrostatics near the gate to drain region of the device changes due to the hot spot created in the device from self heating. This produces an electric field in the direction of accelerating the electrons from the channel to surface states. This will aid to the current collapse phenomenon in the device. Thus, the electric field in the gate to drain region is very critical for reliable performance of the device. Simulations emulating the charging of the surface states were also performed and matched well with experimental data. Methods to improve the reliability performance of the device were also investigated in this work. A shield electrode biased at source potential was used to reduce the electric field in the gate to drain extension region. The hot spot position was moved away from the critical gate to drain region towards the drain as the shield electrode length and dielectric thickness were being altered.
ContributorsPadmanabhan, Balaji (Author) / Vasileska, Dragica (Thesis advisor) / Goodnick, Stephen M (Committee member) / Alford, Terry L. (Committee member) / Venkatraman, Prasad (Committee member) / Arizona State University (Publisher)
Created2013
Description
Since its inception about three decades ago, silicon on insulator (SOI) technology has come a long way to be included in the microelectronics roadmap. Earlier, scientists and engineers focused on ways to increase the microprocessor clock frequency and speed. Today, with smart phones and tablets gaining popularity, power consumption has become a major factor. In this thesis, self-heating effects in a 25nm fully depleted (FD) SOI device are studied by implementing a 2-D particle based device simulator coupled self-consistently with the energy balance equations for both acoustic and optical phonons. Semi-analytical expressions for acoustic and optical phonon scattering rates (all modes) are derived and evaluated using quadratic dispersion relationships. Moreover, probability distribution functions for the final polar angle after scattering is also computed and the rejection technique is implemented for its selection. Since the temperature profile varies throughout the device, temperature dependent scattering tables are used for the electron transport kernel. The phonon energy balance equations are also modified to account for inelasticity in acoustic phonon scattering for all branches. Results obtained from this simulation help in understanding self-heating and the effects it has on the device characteristics. The temperature profiles in the device show a decreasing trend which can be attributed to the inelastic interaction between the electrons and the acoustic phonons. This is further proven by comparing the temperature plots with the simulation results that assume the elastic and equipartition approximation for acoustic and the Einstein model for optical phonons. Thus, acoustic phonon inelasticity and the quadratic phonon dispersion relationships play a crucial role in studying self-heating effects.
ContributorsGada, Manan Laxmichand (Author) / Vasileska, Dragica (Thesis advisor) / Ferry, David K. (Committee member) / Goodnick, Stephen M (Committee member) / Arizona State University (Publisher)
Created2013
Description
Solar energy, including solar heating, solar architecture, solar thermal electricity and solar photovoltaics, is one of the primary energy sources replacing fossil fuels. Being one of the most important techniques, significant research has been conducted in solar cell efficiency improvement. Simulation of various structures and materials of solar cells provides a deeper understanding of device operation and ways to improve their efficiency. Over the last two decades, polycrystalline thin-film Cadmium-Sulfide and Cadmium-Telluride (CdS/CdTe) solar cells fabricated on glass substrates have been considered as one of the most promising candidate in the photovoltaic technologies, for their similar efficiency and low costs when compared to traditional silicon-based solar cells. In this work a fast one dimensional time-dependent/steady-state drift-diffusion simulator, accelerated by adaptive non-uniform mesh and automatic time-step control, for modeling solar cells has been developed and has been used to simulate a CdS/CdTe solar cell. These models are used to reproduce transients of carrier transport in response to step-function signals of different bias and varied light intensity. The time-step control models are also used to help convergence in steady-state simulations where constrained material constants, such as carrier lifetimes in the order of nanosecond and carrier mobility in the order of 100 cm2/Vs, must be applied.
ContributorsGuo, Da (Author) / Vasileska, Dragica (Thesis advisor) / Goodnick, Stephen M (Committee member) / Sankin, Igor (Committee member) / Arizona State University (Publisher)
Created2013
Description
ABSTRACT The D flip flop acts as a sequencing element while designing any pipelined system. Radiation Hardening by Design (RHBD) allows hardened circuits to be fabricated on commercially available CMOS manufacturing process. Recently, single event transients (SET's) have become as important as single event upset (SEU) in radiation hardened high speed digital designs. A novel temporal pulse based RHBD flip-flop design is presented. Temporally delayed pulses produced by a radiation hardened pulse generator design samples the data in three redundant pulse latches. The proposed RHBD flip-flop has been statistically designed and fabricated on 90 nm TSMC LP process. Detailed simulations of the flip-flop operation in both normal and radiation environments are presented. Spatial separation of critical nodes for the physical design of the flip-flop is carried out for mitigating multi-node charge collection upsets. The proposed flip-flop is also used in commercial CAD flows for high performance chip designs. The proposed flip-flop is used in the design and auto-place-route (APR) of an advanced encryption system and the metrics analyzed.
ContributorsKumar, Sushil (Author) / Clark, Lawrence (Thesis advisor) / Bakkaloglu, Bertan (Committee member) / Ogras, Umit Y. (Committee member) / Arizona State University (Publisher)
Created2014
Description
Switch mode DC/DC converters are suited for battery powered applications, due to their high efficiency, which help in conserving the battery lifetime. Fixed Frequency PWM based converters, which are generally used for these applications offer good voltage regulation, low ripple and excellent efficiency at high load currents. However at light load currents, fixed frequency PWM converters suffer from poor efficiencies The PFM control offers higher efficiency at light loads at the cost of a higher ripple. The PWM has a poor efficiency at light loads but good voltage ripple characteristics, due to a high switching frequency. To get the best of both control modes, both loops are used together with the control switched from one loop to another based on the load current. Such architectures are referred to as hybrid converters. While transition from PFM to PWM loop can be made by estimating the average load current, transition from PFM to PWM requires voltage or peak current sensing. This theses implements a hysteretic PFM solution for a synchronous buck converter with external MOSFET's, to achieve efficiencies of about 80% at light loads. As the PFM loop operates independently of the PWM loop, a transition circuit for automatically transitioning from PFM to PWM is implemented. The transition circuit is implemented digitally without needing any external voltage or current sensing circuit.
ContributorsVivek, Parasuram (Author) / Bakkaloglu, Bertan (Thesis advisor) / Ogras, Umit Y. (Committee member) / Song, Hongjiang (Committee member) / Arizona State University (Publisher)
Created2014
Description
In this work, experimental photonic power converter (PPC) design, fabrication and characterization has been used, along with electrical and optical modeling, to study theoretical efficiency limits of monochromatic photovoltaic (PV) energy conversion due to photon recycling. The back-surface reflectance of a photovoltaic (PV) cell is known to strongly influence external radiative efficiency, a photon recycling metric (ERE), especially when reflectance is close to 100 %. Considering a perfect back reflector, an upper PV cell efficiency limit of 70.9 % and 85 % is calculated for 870.7 nm illumination at an intensity that would generate 32 mA/cm2 (1-sun) and 100 A/cm2 (3125-sun eq) photocurrent, respectively. However, when realistic non-idealities are introduced, ideal efficiency can drop by 21 % for both cases as long as the series resistivity for cells under high intensity illumination is limited to 1 mΩ cm^2. This presents a challenge for photonic energy conversion technology where high intensity lasers are typically used to deliver power to equipment from remote locations. This work discusses ways to provide reflectance enhancement while allowing sufficient current flow at the back surface. One way to do this is to use a planar transparent conductive oxide and reflective metal at the back surface. This work measures and compares the back-surface reflectance of IZO/Ag to standard reflective/conductive materials such as Au and Ag. A comparison between cells with the highest V_OC for cells processed with Au and IZO/Ag as reflective back contacts show that the V_OC for the IZO/Ag cell outperforms that of the Au cell by 6.6 mV measuring V_OC=1.071 V with a cell efficiency of 51.0 % at 780 nm LED illumination. Efficiency calculations extrapolated to other monochromatic light sources identified 841 nm as the optimal wavelength for the IZO/Ag cells with a projected efficiency of η_cell=55.5 % for incident intensity corresponding to 1-sun photocurrent. With the fill factors comparable between the cell types, at least at intensities near 1-sun equivalent photocurrent, the IZO/Ag reflective back contact design demonstrates benefits from photon recycling while not sacrificing voltage drop due to series resistance compared to cells with a standard Au back contact.
ContributorsBabcock, Sean Joseph (Author) / King, Richard R (Thesis advisor) / Honsberg, Christiana B (Committee member) / Goryll, Michael (Committee member) / Goodnick, Stephen M (Committee member) / Arizona State University (Publisher)
Created2022
Description
Deep neural networks (DNNs), as a main-stream algorithm for various AI tasks, achieve higher accuracy at the cost of increased computational complexity and model size, posing great challenges to hardware platforms. This dissertation first tackles the design challenges of resistive random-access-memory (RRAM) based in-memory computing (IMC) architectures. A new metric, model stability from the loss landscape, is proposed to help shed light on accuracy under variations and model compression and guide a novel variation-aware training (VAT) solution. The proposed method effectively improves post-mapping accuracy of multiple datasets. Next, a hybrid RRAM/SRAM IMC DNN inference accelerator is developed, that integrates an RRAM-based IMC macro, a reconfigurable SRAM-based multiply-accumulate (MAC) macro, and a programmable shifter. The hybrid IMC accelerator fully recovers the inference accuracy post the mapping. Furthermore, this dissertation researches on architectural optimizations for high IMC utilization, low on-chip communication cost, and low energy-delay product (EDP), including on-chip interconnect design, PE array utilization, and tile-to-router mapping and scheduling. The optimal choice of on-chip interconnect results in up to 6x improvement in energy-delay-area product for RRAM IMC architectures. Furthermore, the PE and NoC optimizations show up to 62% improvement in PE utilization, 78% reduction in area, and 78% lower energy-area product for a wide range of modern DNNs. Finally, this dissertation proposes a novel chiplet-based IMC benchmarking simulator, SIAM, and a heterogeneous chiplet IMC architecture to address the limitations of a monolithic DNN accelerator. SIAM utilizes model-based and cycle-accurate simulation to provide a scalable and flexible architecture. SIAM is calibrated against a published silicon result, SIMBA, from Nvidia. The heterogeneous architecture utilizes a custom mapping with a bank of big and little chiplets, and a hybrid network-on-package (NoP) to optimize the utilization, interconnect bandwidth, and energy efficiency. The proposed big-little chiplet-based RRAM IMC architecture significantly improves energy efficiency at lower area, compared to conventional GPUs. In summary, this dissertation comprehensively investigates novel methods that encompass device, circuits, architecture, packaging, and algorithm to design scalable high-performance and energy-efficient IMC architectures.
ContributorsKrishnan, Gokul (Author) / Cao, Yu (Thesis advisor) / Seo, Jae-Sun (Committee member) / Chakrabarti, Chaitali (Committee member) / Ogras, Umit Y. (Committee member) / Arizona State University (Publisher)
Created2022
Description
The goal of this research work is to develop an understanding as well as modelling thermal effects in Si based nano-scale devices using a multiscale simulator tool. This tool has been developed within the research group at Arizona State University led by Professor Dr. Dragica Vasileska. Another research group, headed by Professor Dr. Thornton, also at Arizona State University, provided support with software tools, by not only laying out the physical experimental device, but also provided experimental data to verify the correctness and accuracy of the developed simulation tool. The tool consists of three separate but conjoined modules at different scales of representation. 1) A particle based, ensemble Monte Carlo (MC) simulation tool, which, in the long-time (electronic motion) limit, solves the Boltzmann transport equation (BTE) for electrons, coupled with an iterative solution to a two-dimensional (2D) Poisson’s equation, at the base device level. 2) Another device level thermal modeling tool which solves the energy balance equations accounting for carrier-phonon and phonon-phonon interactions and is integrated with the MC tool. 3) Lastly, a commercial technology computer aided design (TCAD) software, Silvaco is employed to incorporate the results from the above two tools to a circuit level, common-source dual-transistor circuit, where one of the devices acts a heater and the other as a sensor, to study the impacts of thermal heating. The results from this tool are fed back to the previous device level tools to iterate on, until a stable, unified electro-thermal equilibrium/result is obtained. This coupled electro-thermal approach was originally developed for an individual n-channel MOSFET (NMOS) device by Prof. Katerina Raleva and was extended to allow for multiple devices in tandem, thereby providing a platform for better and more accurate modeling of device behavior, analyzing circuit performance, and understanding thermal effects. Simulating this dual device circuit and analyzing the extracted voltage transfer and output characteristics verifies the efficacy of this methodology as the results obtained from this multi-scale, electro-thermal simulator tool, are found to be in good general agreement with the experimental data.
ContributorsQazi, Suleman Sami (Author) / Vasileska, Dragica (Thesis advisor) / Goodnick, Stephen M (Committee member) / Thornton, Trevor J (Committee member) / Ferry, David K (Committee member) / Arizona State University (Publisher)
Created2021